Skip to content

Quick Start

Getting Started

This project utilizes a hybrid setup combining Conda and uv to ensure a stable and reproducible environment, especially for dependencies with complex system-level requirements.

  • Conda is used to bootstrap a base environment with a specific Python version and to manage system-level dependencies like MPI (via mpi4py).
  • uv, a high-performance Python package manager, is used to create an isolated project-local virtual environment (.venv) and manage all Python packages listed in pyproject.toml.

Prerequisites

Before you begin, ensure you have the following installed:

  1. Conda: Miniconda (recommended) or Anaconda.
  2. uv: If you don't have uv installed, run the appropriate command for your OS:
    • macOS / Linux: 
      curl -LsSf https://astral.sh/uv/install.sh | sh
    • Windows (PowerShell): 
      irm https://astral.sh/uv/install.ps1 | iex

Installation Steps

Follow these steps carefully in your terminal to set up the development environment.

Step 1: Create the Conda Base Environment

This "bootstrap" environment provides the core Python interpreter and MPI support.

# 1. Create a Conda environment named "tongsim_env" with Python 3.12.
#    The "-y" flag automatically confirms the installation.
conda create -n tongsim_env python=3.12 -y

# 2. Activate the new environment.
conda activate tongsim_env

# 3. Install mpi4py.
#    This correctly installs both the Python package and its underlying MPI libraries.
conda install mpi4py -y

Note: Your terminal prompt should now be prefixed with (tongsim_env).

Step 2: Create and Sync the Project Virtual Environment

With the Conda environment active, we'll now use uv to build the final, isolated project environment inside the .venv directory.

Make sure you are in the project's root directory (where pyproject.toml is located).

# `uv` will automatically detect the Python from the active Conda environment.
# It will create a new virtual environment at `./.venv` and install all
# dependencies from the "dev" and "multi_agent" groups into it.
uv sync --group dev --group multi_agent

After this command completes, all required Python packages will be installed in the local .venv folder.

Step 3: Deactivate the Conda Base Environment

The bootstrap environment has served its purpose. You can now deactivate it.

conda deactivate

The (tongsim_env) prefix will disappear from your terminal prompt. If you are returned to the (base) environment, you can run conda deactivate again to exit. The setup is now complete.

Daily Workflow

Whenever you want to work on this project, you only need to perform one step: activate the local virtual environment.

  • Windows (Command Prompt): 

    .\.venv\Scripts\activate.bat

  • Windows (PowerShell): 

    .\.venv\Scripts\Activate.ps1

  • macOS / Linux (Bash/Zsh): 

    source .venv/bin/activate

Generate Protobuf Files

After activating the environment, you need to generate the Python code from the Protobuf definitions:

uv run python scripts/generate_pb2.py

Verification

To verify your installation, you can run a pre-trained model. Before executing the command, make sure the Unreal Editor already has the L_MACS main level open:

uv run examples/marl/example/train.py  --config examples/marl/example/config/mappo.yaml --test --method mappo